Cellulose microfibrils are the major structural component of plant secondary cell walls. Their arrangement in plant primary cell walls, and its consequent influence on cell expansion and cellular morphology, is directed by cortical microtubules; cylindrical protein filaments composed of heterodimers of α- and β-tubulin. In secondary cell walls of woody plant stems the orientation of cellulose microfibrils influences the strength and flexibility of wood, providing the physical support that has been instrumental in vascular plant colonization of the troposphere. Here we show that a Eucalyptus grandisβ-tubulin gene (EgrTUB1) is involved in determining the orientation of cellulose microfibrils in plant secondary fibre cell walls. This finding is based on RNA expression studies in mature trees, where we identified and isolated EgrTUB1 as a candidate for association with wood-fibre formation, and on the analysis of somatically derived transgenic wood sectors in Eucalyptus. We show that cellulose microfibril angle (MFA) is correlated with EgrTUB1 expression, and that MFA was significantly altered as a consequence of stable transformation with EgrTUB1. Our findings present an important step towards the production of fibres with altered tensile strength, stiffness and elastic properties, and shed light on one of the molecular mechanisms that has enabled trees to dominate terrestrial ecosystems.
The arrangement of cellulose microfibrils in plant primary cell walls is directed by cortical microtubules; cylindrical protein filaments composed of heterodimers of α- and β-tubulin (Paredez et al., 2006). In secondary cell walls of woody plant stems the orientation of cellulose microfibrils influences the strength and flexibility of wood, and therefore its suitability for fuel, structural timber or pulp for paper.
Thousands of genes expressed during wood formation have been identified (Allona et al., 1998; Bossinger and Leitch, 2000; Paux et al., 2004; Sterky et al., 1998; Tuskan et al., 2006). The majority of these have not been functionally assessed, as many wood traits can only be studied in wood-forming tissues, and progress is slow because of the inherent variability of wood properties, the late age expression of many wood traits as well as difficulties and experimental timeframes associated with the production and analysis of transgenic tree, wood and cell characteristics.
In Arabidopsis thaliana tubulins are encoded by at least six expressed α-tubulin genes and nine expressed β-tubulin genes, which show high levels of sequence homology and specific spatial and temporal expression patterns (Kopczak et al., 1992; Snustad et al., 1992). Microtubule formation begins with the creation of an assembly-competent α/β heterodimer, in which the association of α- and β-subunits is assisted by folding co-factors that direct their binding (Kirik et al., 2002). These assembly-competent α- and β-tubulins can then be incorporated into microtubule arrays. Owing to their diversity of roles, microtubules are constantly rearranged between a growing or shrinking phase with the aid of stabilizing or destabilizing proteins (Heald and Nogales, 2002). This behaviour is influenced by spatial and temporal queues that affect concentrations of both the α- and β-tubulin subunits within the cytoplasm (Heald and Nogales, 2002; Wasteneys, 2002).
Microfibril orientation in the large S2 layer of secondary fibre cell walls in trees is an important wood quality trait, determining fibre and wood stiffness or elasticity, and is simply referred to as the microfibril angle (MFA). It is measured as the microfibril deviation in this cell-wall layer from the long axis of the cell (Long et al., 2000) (Figure 1d,e). Cells with more longitudinally aligned microfibrils have a low MFA and can more readily resist high tension forces. Wood forming in the upper sides of eucalypt branches (tension wood) has a much lower MFA (∼15°) than wood forming in the lower sides (∼40°) (Washusen et al., 2005).
In plant cells undergoing morphogenesis, cortical microtubules are believed to guide microfibril deposition, by directing cellulose synthase rosettes, and thereby determine the orientation of cellulose microfibrils in primary cell walls (Paredez et al., 2006). Histological studies show co-orientation between microtubules and the helical alignment of microfibrils, again during the early stages of secondary cell-wall deposition, suggesting that these processes are also linked (Abe et al., 1995; Chaffey et al., 2002). More recently, cortical microtubules have been shown to influence secondary cell-wall development in cell suspension-induced tracheary elements of A. thaliana (Oda et al., 2005).
To date, two genes have been found to affect microfibril orientation in the secondary cell wall. FRA1, coding for a kinesin-like protein with an N-terminal microtubule binding motor domain, influences the mechanical strength of fibres in Arabidopsis inflorescence stems, and was proposed to be involved in microtubule control of cellulose microfibril order (Zhong et al., 2002). Whether or not FRA1 also has a role in secondary fibre wall formation during xylogenesis in stems of forest trees is unknown. The second gene encodes Cinnamoyl CoA reductase (CCR), a key enzyme in the lignin biosynthetic pathway (Piquemal et al., 1998). Allelic variation in CCR has been correlated with variation in MFA in Eucalyptus using association mapping (Thumma et al., 2005); however, no assessment of specific gene function was attempted. Lignin content and MFA are known to both be inversely correlated with cellulose content (Plomion et al., 2001), and the observed association between allelic variation for CCR and MFA might well be indirect.
Consequently, the molecular mechanisms that direct microfibril orientation, and determine MFA in the secondary fibre cell wall, remain largely unknown. Here we report on how a Eucalyptus grandisβ-tubulin gene (EgrTUB1) was identified as a candidate for association with wood-fibre formation, and how we have functionally linked EgrTUB1 expression with cellulose microfibril orientation through analysis of somatically derived transgenic wood sectors in stems of Eucalyptus globulus trees (Figure 1).
β-tubulin expression in mature trees, gene isolation and phylogeny
A β-tubulin isotype (EgrTUB1) was identified in a differential screen of an E. grandis xylem cDNA library based on its abundance and very strong expression in developing secondary xylem (Figure 2a). Increased EgrTUB1 expression in xylem developing on the upper side of branches (by approximately threefold with respect to the underside of the branch and twofold with respect to main-stem xylem; Figure 2a and microarray data not shown) coincided with wood fibre cells with a lower MFA. Five β-tubulin genes were cloned and sequenced and, using the predicted amino acid sequences for these genes, a phylogenetic tree of plant β-tubulins was generated. EgrTUB1 was distinct from most other plant β-tubulins, clustering on a branch that includes two cotton proteins (Figure 2b).
TUB1 promoter activity in tobacco
Eucalyptus TUB1 promoter activity was evident during xylogenesis in developing wood of transgenic tobacco plants expressing a Eucalyptus nitens TUB1 promoter:GUS fusion construct (Figure 3). There was little to no activity in the cambium and adjacent cells, indicating that EnTUB1 expression is activated during later stages of xylem differentiation that include the deposition of the S2 secondary cell wall. TUB1 activity was also observed in leaf vasculature during seedling growth (from 2.5 weeks post-germination), and to a very minor extent in parts of open flowers; little promoter activity was visible in any other plant tissues throughout development.
Induced somatic sector analysis (ISSA)
Induced somatic sector analysis (Spokevicius et al., 2006; Van Beveren et al., 2006) was used to create and analyse transgenic wood sectors. In this method in vivo transformation of cambial cells in tree stems involves exposing a window of cambial cells in the growing stem via partial bark peeling. A suspension of Agrobacterium tumefaciens carrying a suitable vector that contains a marker gene (e.g. uidA; GUS), together with the target gene to be tested, is applied to the exposed tissue before the peeled bark is moved back into place and cambial growth resumes. Further growth of individual stably transformed cambial initials leads to the formation of transformed wood and phloem sectors, in which candidate gene effects can be studied, in relation to neighbouring untransformed tissue within the same tree, at identical developmental stages and under the same treatment conditions (Figure. 1). The expectation is that target gene effects will only be visible in the GUS-staining sector so long as the candidate gene acts in a cell-autonomous manner.
Using control constructs that only carry CaMV35S:GUS, in vivo transformation efficiency was determined in poplar and eucalypts with the A. tumefaciens strain AGL1. This was measured as the number of transformed tissue sectors created per cm2 of originally exposed cambial tissue. Results of these experiments have been published previously (Van Beveren et al., 2006) and show that stable transformation efficiency in both species is high. In contrast, GUS-staining stem sectors transformed with a construct containing, on the same vector but in different cassettes, the reporter gene and CaMV35S:EgrTUB1 were found only in E. globulus, with no transformed cells identified in Populus alba. Transformed sectors were found in all tissue types, and their respective frequencies are presented in Table 1. Only sectors that originated from stable transformation of a single cambial initial (cambial sectors; i.e. sectors that crossed the cambium and included both newly formed phloem and xylem) were considered for further analysis.
Table 1. Inoculation area and transformation frequencies following in vivo stem transformation with Eucalyptus grandis β-tubulin gene
‘Cambial window’ area (cm2)
Transformed sectors per cm2
Only trials where successful transformation occurred are included. Figures in bold brackets indicate SD and figures in italics indicate minimum and maximum values. Only cambial sectors (i.e. sectors that extended through the cambium both into the phloem and xylem; cf. Figure 1) were used for microfibril angle analyses.
Eucalyptus globulus 1 (Eg1)
E. globulus 2 (Eg2)
Assessment of MFA in and around transgenic wood sectors
The MFA in fibres sampled from sectors transformed with constructs containing either the GUS reporter gene and EgrTUB1 or just GUS were measured, and results were compared with those obtained from adjacent, untransformed control regions (Figure 1b,c). Three separate assessments of MFA were made. Firstly, in order to determine if GUS (or GUS:GFP) had any effect on S2 MFA in either poplar or eucalypts, five independent CaMV35S:GUS-transformed cambial sectors from eucalypts and five CaMV35S:GUS:GFP poplar sectors were dissected, and tissue from cambium region sampling points was processed for MFA determination according to the procedure described in Figure 1 (i.e. 30 fibres were collected from each sampling point, three measurements were taken for each fibre and results were pooled from five sectors for each species). Mean MFA in fibres sourced from GUS-positive (transformed) sectors in either poplar or eucalypts did not differ significantly from adjacent untransformed control regions (0.93°, P = 0.193 for poplar; 0.02°, P = 0.989 for eucalypts).
Secondly, fibres were sourced from nine sampling points (Figure 1b,c) in and around a single EgrTUB1-transformed cambial sector (Eg1CS1; Figure 4). The transformed region consistently showed higher MFA across all sampling zones, with each sampling zone (wound, middle, cambium) representing a distinct developmental stage of cambium recovery after bark peeling. MFA in all sampling points was found to be higher in the middle and wound zones (presumably in response to wounding) when compared with the cambium zone (Figure 4). These results demonstrate the variability of MFA depending on developmental stage, highlighting the importance of comparing MFA in transformed, GUS-staining sectors with control tissue sourced in the same sampling zone (i.e. at the same developmental stage; Figure 4). In this assessment the cambium zone was determined to be most representative of MFA values expected in comparable non-wounded eucalypt stems (Washusen et al., 2005).
The third assessment of MFA used fibres from the three sampling regions in the cambium sampling zone only. In this assessment, fibre MFA in and around a further four individually transformed cambial EgrTUB1 sectors (Eg2CS2–5) was assessed. Median values for each sampling point were calculated and grouped according to sampling region for statistical analysis, which revealed that MFA in the transformed sampling region was significantly higher than in both control regions (Near 4.97°; P = 0.00036; Far 5.43°; P = 0.0002; Figure 5). The control regions showed no significant differences when compared with each other (0.46°; P = 0.599; Figure 5).
EgrTUB1 is involved in secondary wall formation during eucalypt xylogenesis
The relatively high number of β-tubulin genes identified in plants (at least nine have been identified in Arabidopsis) (Snustad et al., 1992) and the phylogenetic diversity of their deduced predicted amino acid sequences (Figure 2b) reflect their diversity of roles during plant morphogenesis. In our experiments, tissue-specific expression of EgrTUB1 is evident from Northern blot experiments, which show very high expression in xylem tissue and weak expression in the phloem (Figure 2a). Tissue and developmental-stage specificity is further supported by our promoter studies in tobacco, where TUB1 promoter-driven GUS expression is clearly observed in the vasculature of young leaves (Figure 3b), and in developing xylem tissue in the stem, but is absent from the cambial region (Figure 3f). These results suggest a role for EgrTUB1 during secondary wall formation and, together with the distinctive phylogenetic position of EgrTUB1 (Figure 2b), indicate a specific role for this protein during wood formation in Eucalyptus.
Upregulation of β-tubulin in poplar is cytotoxic
The absence of CaMV35S:EgrTUB1-sectors in P. alba was of particular interest. P. alba stem tissue is highly susceptible to in vivo transformation (Spokevicius et al., 2006; Van Beveren et al., 2006), suggesting EgrTUB1 overexpression was lethal. Support for this interpretation comes from experiments in maize, where transformed tissue in which only a single tubulin subunit within a heterodimer was overexpressed did not regenerate (Anthony and Hussey, 1998). In maize (Anthony and Hussey, 1998) and tobacco (Anthony et al., 1999) tissue regeneration occurred only when both tubulin heterodimer subunits were overexpressed in combination. Lack of regeneration resulted from plant cells being unable to tolerate large imbalances in the ratio of α- to β-tubulin when constitutive overexpression of only one subunit occurred. Using an anti-sense approach, full plants were able to regenerate, where only the α-tubulin subunit was targeted, but only small decreases in α-tubulin were detected in regenerated plants (Bao et al., 2001). Also, in these studies a decrease in the regeneration efficiency was noted, and the authors suggest that in cells in which the silencing effect was high the imbalance was likely to have caused cell death, and therefore reduced transformation and regeneration efficiency.
EgrTUB1 expression in transgenic wood sectors in Eucalyptus stems is possibly decreased because of homology-dependent gene silencing (HDGS)
It seems surprising that in E. globulus a high number of EgrTUB1-transgenic sectors were recorded (Table 1). This could point to the existence of a special mechanism in Eucalyptus by which cells can reduce the cytotoxic effects of β-tubulin. There is evidence in Arabidopsis that tubulin-folding co-factor A, a microtubule chaperone protein, is involved in maintaining a correct balance of α- and β-tubulins by binding and consequently reducing the toxic effects of free β-tubulins (Kirik et al., 2002). It seems unlikely, however, that a naturally occurring balancing mechanism, capable of accommodating large increases in β-tubulin, would occur in eucalypts but not in poplar.
A more plausible explanation is that expression of β-tubulin did not increase in these sectors, but rather decreased as a result of HDGS (van der Krol et al., 1990; Napoli et al., 1990). HDGS in plants is triggered by high sequence homology between endogenous and exogenous gene sequences, leading to variable decreases in the expression of both homologues. This phenomenon has been observed in many plant species including trees (Baucher et al., 1996; Tsai et al., 1998). Although the EgrTUB1 nucleotide sequence is conserved in eucalypts (data not shown), it differs substantially from any potential P. alba homologue (Figure 2b). Homology between EgrTUB1 and the endogenous E. globulus TUB1 gene would have been sufficient to induce HDGS in the transgenic tissue in E. globulus, whereas the lack of homology in P. alba enabled overexpression and death of the transformed cells. Consequently, we argue that no sectors were observed in P. alba as a result of the cytotoxic effect of the higher proportion of β-tubulin, and sectors were recovered in E. globulus because of (mild) suppression of β-tubulin via HDGS. Confirmation of this hypothesis is not yet possible as the detection of transgenic wood sectors using GUS assays is destructive, and consequently quantitative gene expression studies in such sectors remain a technical limitation.
MFA in Eucalyptus is correlated with EgrTUB1 expression
In Eucalyptus, downregulation of EgrTUB1 was shown to be associated with higher MFA, whereas upregulation was associated with lower MFA on the lower and upper sides of branches, respectively (compare Figure 2a and Washusen et al., 2005). Consequently, overexpression of EgrTUB1 in xylogenic tissue, provided it did not lead to dramatic loss of cell viability, was predicted to cause a lower average MFA, whereas downregulation was predicted to cause a higher average MFA. In our experiments a significant increase of approximately 5° mean MFA in transgenic fibres was observed (Figure 5), further supporting the hypothesis that downregulation occurred.
β-tubulin determines cellulose microfibril orientation during xylogenesis
Microtubules have been implicated with guiding microfibril deposition in secondary cell walls through establishment of a scaffold in the plasma membrane (Baskin, 2001; Roberts et al., 2004). Our results using differential gene expression studies, promoter studies and ISSA provide functional evidence that β-tubulin (either cytoplasmic or as a structural component of microtubules) is directly involved in the determination of cellulose microfibril orientation during xylogenesis. We speculate that changes in relative proportions of specific β-tubulin monomers in microtubules could influence, possibly via changes to microtubule structure, the positioning of cellulose microfibrils in secondary cell walls. These findings demonstrate that molecular editing can be successfully used for altering a commercially important complex wood trait, representing an important step towards the production of fibres with altered tensile strength, stiffness and elastic properties. Our results also provide a starting point for understanding the molecular mechanisms that have enabled trees to dominate terrestrial ecosystems.
Construct design and A. tumefaciens preparation
The E. grandis β-tubulin1 gene (EgrTUB1) was cloned into the pCAMBIA 1303 binary vector (http://www.cambia.org) with the resulting vector being termed pEGTUB+. A 538-bp CaMV35S PCR fragment, using primers 5′-CCGAATTCATGGAGTCAAAGATTC-3′ and 5′-CCCGGTACCGAGCTCAGTCCCCCGTGTTCTC-3′, and a 253-bp Nopaline Synthase (NOS) terminator PCR fragment using primers 5′-GGCTGCAGGCATGCCGTTCAAACATTTGGC-3′ and 5′-GGGGAAGCTTCCCGATCTAGTAACATAG-3′ were ligated into the KpnI and EcoRI sites, and PstI and HindIII sites, respectively, of the pCAMBIA 1303 binary vector to create p1303:35S:NOS. A full-length EgrTUB1 cDNA PCR fragment amplified using primers 5′-AGTTTCTAGAGCTCAAGATGAG-3′ and 5′-CATACTGCAGTTTCCCCTGTTCAATC-3′ was ligated into the XbaI and PstI sites of p1303:35S:NOS to yield pEGTUB+. The p1303:35S:NOS construct (without the addition of EgrTUB1) as well as pCAMBIA 1305.1 (featuring the GUS Plus reporter gene) were used for the creation of control sectors.
The EnTUB1 promoter was isolated from an E. nitens Genomewalker library (constructed using a Genomewalker Universal Kit; Clontech, http://www.clontech.com), utilizing primers based on the homologous EgrTUB1 cDNA. A 1013-bp EnTUB1 promoter PCR fragment (using the primers 5′-AAAGGGATGCTCCAACACC-3′ and 5′-CATGCCATGGTGAGCAAGATGAACTAGGCAAG-3′) immediately upstream of the ATG was cloned into the BamHI and NcoI sites of the pCAMBIA1381Z binary vector, thus yielding the EnPTUB1:GUS construct, a fusion between the EnTUB1 promoter and the GUS reporter gene. The LBA4404 strain of A. tumefaciens was transformed with the EnPTUB1:GUS construct using triparental mating.
Transformation of tobacco plants
Nicotiana tabacum W38 (tobacco) leaf discs were transformed with the EnPTUB1:GUS construct via A. tumefaciens, and hygromycin-resistant transformants regenerated into T1 plants. Results are presented from the four strongest expressing T2-generation lines derived from independent T1 transformants. All transgenic plants were confirmed to contain the transgene by PCR using gene-specific primers.
Harvesting and preparation of samples for histological assays
EnTUB1 promoter activity was examined by GUS staining in a growth series of whole tobacco seedlings (4-days old, and 1, 2, 3 and 4 weeks post-germination) and in individual plant tissues (roots, leaves, stems and flowers) of the whole mature plant (4 months). For each T2 line, three or four plants were harvested and GUS stained for each time point according to published protocols (Jefferson et al., 1987).
Plant material used for ISSA
Eucalyptus globulus Labill. seedlings of unknown provenance were purchased from a local nursery and maintained in a shadehouse. Clonal P. alba‘pyramidalis’ L. plants growing at the University of Melbourne Creswick Campus, Victoria, Australia, were either sourced from lingo-tuberous individuals, transplanted from the ground and placed in pots, or from cuttings treated with rooting hormone powder (Yates rooting powder, http://www.yates.com.au) in cutting beds, allowed to grow in a glasshouse for several months, and were then transplanted to pots and placed in a shadehouse. Plants aged between 6 and 12 months were transferred to controlled glasshouse conditions, where temperatures ranged from 14 to 17°C at night and from 21 to 25°C during the day. Extra light was provided during winter months to achieve > 12-h photoperiods. Plants were watered regularly (as required depending on the season) and were fertilized with a slow-release fertilizer (OsmocoteTM, Scotts-Sierra Horticultural Products, http://www.scotts.com) every 6 months.
Constructs and A. tumefaciens preparation for ISSA
Methods for the preparation of A. tumefaciens (AGL1) for ISSA varied. In method 1, used for the in vivo transformation of cambial cells involving E. globulus stems, an individual colony of A. tumefaciens harbouring pEGTUB+ was grown in 4 ml of Luria Bertani broth (LB), containing appropriate antibiotics, for two nights and used directly for inoculation. In method 2, bacteria that were grown over 48 h as described for method 1 were used as a starter culture that was grown to log-phase, and then collected as a pellet. Bacteria were then re-suspended in 5 ml of induction medium [2 mm phosphate buffer, 1x AB salts (from 20x stock: 373 mm NH4Cl, 24 mm MgSO4.7H2O, 40 mm KCl, 1.7 mm CaCl2.H2O, 0.18 mm FeSO4.7H2O), 30 mm 2-(N-morpholine)-ethanesulphonic acid (MES), 0.5% glucose] containing 100 μm of acetosyringone and left shaking gently on an orbital shaker (20–40 rpm) in a growth room (at 21°C) overnight. Bacteria were pelleted (1150 g, 15 min), re-suspended in 1 ml of MS medium and used directly on plants.
In vivo transformation of stem and bud tissue
For transformation of E. globulus with pEGTUB+ two separate trials using in vivo stem transformation methods (Van Beveren et al., 2006) were conducted in plants that had been acclimatized to glasshouse conditions (24°C day, 17°C night, 16-h photoperiod). In the first trial (Eg1), a single 1-cm2 cambial window (‘upward flap’) was exposed using a scalpel on ten separate tree stems. These were each inoculated with 10 μl of A. tumefaciens suspension prepared using method 1 and the phloem strip replaced and wrapped firmly with ParafilmTM (Alcan Packaging, Neenah, WI, USA; http://www.parafilm.com) between five and eight times. In the second trial (Eg2) this protocol was altered slightly. Two separate 1-cm2 cambial windows were created on nine separate plants and 10 μl of A. tumefaciens prepared using method 2 was introduced. Here two extra vertical cuts were made through the phloem strip, which was then replaced and wrapped firmly with ParafilmTM between five and eight times. Plants were watered well, fertilized with OsmocoteTM when necessary, and left to grow for either eight (second trial) or eleven months (first trial) prior to harvesting.
For P. alba one trial, involving in vivo bud (Spokevicius et al., 2006) and in vivo stem (Van Beveren et al., 2006) methods, was conducted. Firstly, dormant lignotuber plants with robust bud set were moved into the glasshouse, placed in potting mix and allowed to acclimatize for 2 days. Glasshouse conditions were conducive to bud burst (24°C day, 17°C night, > 16-h photoperiod) and for experimentation 341 dormant lateral buds on 19 plants were severed, inoculated and wrapped in ParafilmTM according to the methods described earlier (Spokevicius et al., 2006). After 2 months, seven of these 19 plants that showed good growth were used for further in vivo stem transformation. Using the same protocols as described for the Eg2 trial, six 1-cm2 cambial windows (42 in total) were created in each of the seven P. alba plants. All plants were well watered, fertilized when necessary and left to grow for a further 5 months prior to harvest.
For the creation of control sectors the in vivo stem method was used. Plants and A. tumefaciens were prepared as described previously, and 30 1-cm2 windows on 10 plants each (three windows per plant) were inoculated with the p1303:35S:NOS (poplar) or pCAMBIA1305.1 (eucalypt) constructs. Plants were well watered, fertilized when necessary and left to grow for a further 3 months prior to harvest.
Harvesting and preparation of ISSA samples for histological assays
Stem tissue from in vivo stem and in vivo bud trails was harvested by excising either the complete cambial window or the transformed stem area and then by further dissection. For in vivo stem derived tissue, samples where dissected longitudinally into rectangular blocks between 0.5- and 2-mm thick and of 5–10 mm in width and length. For in vivo bud derived tissue, samples were dissected longitudinally from the pith to bark into pieces of pie between 0.5- and 2-mm thick and of 5–10 mm in width and length. In both cases this was performed to preserve the radial and tangential fibre walls for later phenotypic assessment. GUS histological assaying, storage and assessment protocols have been described previously (Spokevicius et al., 2006; Van Beveren et al., 2006). Growth data was collected, including plant height and diameter at 10 cm for plants involved in the in vivo stem trials, and plant diameter and individual shoot length and diameter at 2 mm in plants in the in vivo bud trials.
Fibre preparation for MFA assessment
The protocol used to assess wood fibre MFA involved an indirect method where pit apertures on the longitudinal cell walls of fibres were assessed. This method has been used successfully to deduce MFA by measuring the angle of the long axis of the ellipse in pit apertures relative to the longitudinal axis of the fibre. Although not present on all fibres measured, cell-wall striations indicative of MFA alignment were observed and taken into account during measurement (Figure 1d,e). Fibres were sampled from one of nine sampling points taken from a sampling region (transformed, near, far) within a sampling zone (cambium, middle, wound). Sampling points are depicted in Figure 1b,c. In all cases the location of the near sampling region was 1–2 mm away from the transformed region, and the far sampling region was 4–5 mm away from the transformed region.
Fibres from sampling points were excised as small blocks no larger than 1 mm3, from wood blocks using a double-edged razor blade under an Olympus SZ-PT stereo microscope (Olympus, Mount Waverley, VIC, Austria; http://www.olympus.com). Care was taken in sampling from the transformed sampling region to ensure that only transformed fibres, or those likely to be transformed fibres (owing to the loss of GUS-staining during maturation and cell death), were taken. In sampling regions outside the transformed area this level of care was not required, and often larger quantities of tissue could be obtained. Excised blocks were placed in separate 1.5-mL tubes containing 1:1 hydrogen peroxide and glacial acetic acid, and were macerated for 2 h at 90°C. After this, fibres were removed and rinsed in distilled water twice, mounted on glass slides with Gelvastab mounting media (Fluka, http://www.sigmaaldrich.com/Brands/Fluka___Riedel_Home.html) and, with sharp tweezers, teased apart so that individual fibres could be visualized. A slide cover was placed on top, light pressure was applied and slides were left overnight to set. Maceration led to a slight loss in GUS staining intensity. Some staining, however, could still be observed after treatment.
The following day, slides were assessed under an Olympus BH-2 light microscope (Olympus, Mount Waverley, VIC, Austria; http://www.olympus.com with the aid of polarized light, and photos of pit apertures on individual fibres were taken using a Leica DC 100 digital camera (Leica, http://www.leica.com) at 59 pixels per cm and processed using im50 software. The polarized light could be rotated over the light source, and helped with distinguishing the long axis of the pit aperture ellipse more clearly and made cell-wall striations more visible. In general, fibres were randomly chosen on slides, except for fibres from the transformed sampling point, where preference was given to measuring fibres that either showed some GUS staining or were in close proximity to stained fibres or rays. A total of 30 fibres were photographed and assessed per sampling point, with the angle of three pit apertures measured per fibre. In reports where MFA was measured using pit apertures, 25–30 fibres per sampling point were shown to be sufficiently informative (Jang, 1998). For statistical analysis, anova (with repeated measurements) and least significant difference (LSD) tests were used to identify significant differences between sampling regions.
Molecular and phylogenetic analysis of plant β-tubulins
RNA isolation from plant tissues and northern analysis were carried out as previously described (Southerton et al., 1998). A probe containing 105 bp of the C-terminal and 176 bp of the 3′ untranslated region (UTR) of EgrTUB1 was amplified using gene-specific primers. The probe gave a single band when hybridized to Southern blots of eucalypt genomic DNA, indicating hybridization at a single locus (data not shown).
The EgrTUB1 full-length cDNA was used as a probe to screen an E. grandis xylem cDNA library, isolating an additional three full-length sequences and one partial β-tubulin cDNA sequence. Predicted amino acid sequences of β-tubulins from A. thaliana, Gossypium hirsutum, Lycopersicum esculentum, Oryza sativa, Populus trichocarpa, Zea mays and E. grandis were aligned using clustalw, and the unrooted dendrogram was obtained using mega 3.1. Codes, accession numbers and gene IDs used in this study include Arabidopsis Genome Initiative codes, GenBank accession numbers, Populus contig IDs and unigene IDs: A. thaliana, AtTUB1 (At1 g75780), AtTUB2 (At5 g62690), AtTUB3 (At5 g62700), AtTUB4 (At5 g44340), AtTUB5 (At1 g20010), AtTUB6 (At5 g12250), AtTUB7 (At2 g29550), AtTUB8 (At5 g23860), AtTUB9 (At4 g20890); G. hirsutum, GhTUB1 (AF487511), GhTUB2 (Ghi.8448), GhTUB3 (Ghi.6127), GhTUB4 (Ghi.1175), GhTUB5 (Ghi.7579), GhTUB6 (Ghi.4629); L. esculentum, LeTUB1 (BT013153), LeTUB2 (BT014148), LeTUB3 (BT012803), LeTUB4 (DQ205342), LeTUB5 (BT013893), LeTUB6 (BT013141); O. sativa, OsTUB1 (Os.46904), OsTUB2 (Os.5233), OsTUB3 (4331315), OsTUB4 (Os.11093), OsTUB5 (4326917), OsTUB6 (Os.45929), OsTUB8 (Os.11155) OsTUB3 (Os.12811); P. trichocarpa, PtTUB1 (POPLAR.85.C15), PtTUB2 (POPLAR.85.C8), PtTUB3 (POPLAR.85.C13), PtTUB4 (POPLAR.85.C1), PtTUB5 (POPLAR.85.C6), PtTUB6 (POPLAR.85.C16); Z. mays, ZmTUB1 (X52878), ZmTUB2 (Zm.79), ZmTUB3 (Zm.7172), ZmTUB4 (Zm.3257), ZmTUB5 (Zm.4718), ZmTUB6 (Zm.7750), ZmTUB7 (Zm.16988), ZmTUB8 (Zm.4718); E. grandis, EgrTUB1 (EF534219), EgrTUB2 (EF534220), EgrTUB3 (EF534221), EgrTUB4 (EF534222), EgrTUB5 (EF534223). Gene names are as published (A. thaliana, O. sativa and Z. mays) or were assigned arbitrarily.
During this work AS was supported by an Australian Post-graduate Award (Industry). DQ and SG were supported by post-doctoral fellowships jointly funded by the Commonwealth Scientific and Industrial Organization (CSIRO) and the Administration of Forestry and Fishery of Australia (AFFA). Much of this work was funded through an Australian Research Council (ARC) Linkage Grant (LP0211919), and ongoing support by Sappi (South Africa) and Arlene Bayley is gratefully acknowledged. We also thank Lawrie Wilson, Kim Van Beveren and Alexander Myburg for many helpful discussions, and Michael Tausz for critically reading our manuscript.